|Page tools: Print Page Print All RSS Search this Product|
UNDERSTANDING EARNINGS IN AUSTRALIA USING ABS STATISTICS
This article explores some of the earnings statistics produced by the ABS, through:
WHAT DO WE MEAN BY EARNINGS?
In the broadest sense, earnings can be thought of as amounts paid by employers to employees for work done. More specifically, earnings are the pre-tax amount paid to employees for work done or time worked (including paid leave). In concept, earnings include 'payments-in-kind' - i.e. the value of 'non-cash' goods or services provided to employees (fringe benefits), however in practice in ABS collections they are not included. Wages and salaries in cash also conceptually includes the value of goods and services obtained through salary sacrifice arrangements, where it is the choice of the employee. For more information on the conceptual framework for employee remuneration see Information paper: Changes to ABS measures of employee remuneration, 2006 (cat. no. 6313.0).
Earnings in ABS statistics are consistent with international definitions determined by the International Labour Organisation and in the System of National Accounts (2008).
For more detailed definitions and descriptions of the concept of earnings, refer to Chapter 12 of Labour Statistics: Concepts, Sources and Methods (cat. no. 6102.0.55.001).
KEY SOURCES OF EARNINGS DATA
The ABS produces earnings statistics, as well as earnings related measures, from a range of sources. The major sources of earning statistics in the ABS, and the publications in which they are released, are:
Household and employer surveys which are used by the ABS to collect earnings statistics have different strengths and limitations. It is important to be aware of these differences when analysing the data.
The rest of this article focusses on three key ABS labour surveys providing estimates of earnings and explains the purpose and key outputs of each, as well as their benefits and limitations. The surveys are:
In addition, the ABS WPI (cat. no. 6345.0), which provides a measure of changes in wages and salaries paid by employers for a unit of labour (i.e. hour) over time, is discussed as movements in WPI are often compared to AWE.
The first two surveys, EEH and AWE are employer surveys and measure earnings related to a 'point in time' (e.g. a pay period). They collect wages and salaries in cash that are received regularly and frequently (e.g. exclude one-off bonuses) and include payments for employees on paid leave.
EEBTUM is a household survey and also collects earnings at a 'point in time', the most recent pay period, i.e. the last total pay. It collects wages and salaries in cash, before tax or any other deductions. As the survey collects amounts of "total last pay", it may include irregular and infrequent payments or bonuses, and payments related to other periods.
Survey of Employee Earnings and Hours
The two-yearly EEH provides statistics on the composition and distribution of employee earnings, the hours paid for, and the methods used to set their pay. From 2006, estimates of earnings from EEH have included amounts salary sacrificed.
The information in EEH is collected from businesses but at the individual employee level. This makes it possible to derive measures of distribution (e.g. medians, deciles, earnings ranges) and provide some information on individual characteristics of employees. The median is a better measure of 'central tendency' than the mean when distributions are uneven or skewed, as the mean can be heavily influenced by outliers in the distribution. This is discussed in more detail later.
EEH also provides some information on individual characteristics of employees. These include: managerial/non-managerial status; occupation; sex; full-time/part-time status; adult/junior status; type of employee (permanent, fixed-term contract or casual); method of setting pay (i.e. award only, collective agreement and individual arrangement); and hours paid for. From 2014 onwards age of employee will also be collected in EEH. The EEH survey therefore complements the AWE survey by providing detailed information on the composition and distribution of employee earnings and hours, however on a less frequent basis.
A key strength of EEH is that it allows for hourly measures of earnings to be derived (currently only for non-managerial employees). Hourly earnings measures are useful for comparisons between groups who may work different weekly hours.
Non-managerial adult hourly ordinary time earnings from EEH is a widely used measure, since it allows as much of a like-for-like comparison as possible, facilitating comparison of earnings for different population groups. For example directly comparing the weekly earnings of full-time and part-time employees would not take hours paid for into account.
Survey of Average Weekly Earnings
The six-monthly AWE is currently the most frequently available source of the level of earnings. It is designed to provide estimates of the level of average earnings at a point in time, and while not designed for movements in earnings, the frequency of collection supports a time series of these level estimates. Data on the average level of earnings are useful for providing a level benchmark to compare a specific amount to an average level of earnings e.g. what an individual earns compared to the average.
AWE has the longest history of the three ABS earnings sources discussed in this article. Collecting average earnings data is relatively simple and can produce estimates in a timely manner. While not designed as an index of wages, it is extensively referenced in legislation for indexation purposes.
Data are obtained from selected businesses on the total earnings (ordinary time and overtime) paid to their employees and the total number of employees in the business, which together are used to derive the mean, or average, earnings. These sample data are then weighted to provide estimates for the whole population of in scope businesses. Estimates are available by state/territory, sex, industry and sector.
The three key earnings series (excluding amounts salary sacrificed) produced from AWE are:
The earnings series from AWE historically excluded amounts salary sacrificed. As discussed above, amounts salary sacrificed are conceptually part of wages and salaries in cash, however, the key earnings series from AWE have continued to be published on the old conceptual basis (i.e. exclusive of amounts salary sacrificed) to maintain long term comparability of the key series. Since the May 2011 AWE publication, the Average Weekly Cash Earnings (AWCE) series have also been released. These series are inclusive of salary sacrificed amounts. For more information see the Explanatory Notes of the AWE publication (cat. no. 6302.0) and Information paper: Changes to average weekly earnings, Australia (cat. no. 6302.0.55.002).
Out of the three series produced from AWE, the AWOTE for full-time adult employees series is generally considered the most stable earnings series due to the exclusion of overtime and part-time and junior employees, however it should be noted that the series does not represent all employees. AWTE for full-time employees has higher levels compared to AWOTE for full-time employees as it includes overtime. AWTE series for all employees has the lowest levels as it includes the earnings of part-time and junior employees, who receive lower pay on average than full-time adult employees.
Compositional changes in the employee population (e.g. the mix between full-time and part-time employees, or the industries and/or occupations in which they work) and the composition of the survey samples selected, can impact on the level of average earnings. For example, if there is an increase in part-time employment then, all other things being equal, the average weekly total earnings series for all employees would be expected to decrease.
Survey of Employee Earnings, Benefits and Trade Union Membership
EEBTUM is a household survey, conducted annually as a supplement to the monthly Labour Force Survey (LFS). This survey collects weekly earnings data together with a range of socio-demographic information collected from individual people, such as: sex; age; marital status; relationship in household; geographic region of usual residence; school attendance; country of birth; and year of arrival in Australia.
EEBTUM also collects details about the nature of employment, including: occupation; industry; hours worked (hours paid for, hours actually worked and hours usually worked); full-time/part-time status based on hours worked; sector; size of workplace; and leave entitlements. From 2007, EEBTUM has included amounts salary sacrificed in the estimates of earnings.
As EEBTUM is collected at the individual employee level, like the EEH survey, this means that measures of earnings distribution (e.g. medians, deciles, earnings ranges) are able to be produced.
The three surveys discussed above have important differences in concepts, scope and methodology, which can result in different estimates of weekly earnings. Therefore, care should be taken when comparing estimates of earnings from these surveys. The main differences are described in the box below.
Wage Price Index
The WPI measures changes in wages and salaries paid by employers for a unit (i.e. hour) of labour where the quality and quantity of labour are held constant. It is widely used as a measure of wage and salary inflation in the economy.
While AWE provides estimates of the level of earnings at a point in time, the quarterly WPI is a more relevant indicator for changes in the rates of pay. For further information on the WPI, please refer to the Explanatory Notes of Wage Price Index, Australia (cat. no. 6345.0) and Wage Price Index: Concepts, Sources and Methods (cat. no. 6351.0.55.001).
Period-to-period movements for the AWE series are not necessarily comparable with those for the WPI. It is important to recognise that the two series have different purposes and concepts, and use different sample selection, rotation, and estimation methodologies.
The WPI measures change in the price employers pay for labour that arise from market factors. Specifically, the WPI measures change in the price of wages and salaries. As a price index the quantity and quality of labour services are held constant, changes in the composition of the labour force, hours worked, or changes in characteristics of employees (e.g. work performance) are all excluded from the index. For the WPI this is achieved by ensuring that identical jobs are priced from one period to the next. This is referred to as pricing to constant quality.
USES OF EARNINGS DATA
Earnings statistics provide information on both the levels and movements in average earnings, and on the distribution of earnings for different groups of employees. Earnings statistics available from ABS sources provide key indicators to help inform policy, research and discussions of important labour market issues such as pay equity, social welfare, wage setting and income distribution. It is important to understand the relative strengths and limitations of the various earnings sources to ensure appropriate interpretation of the statistics.
As discussed above, there are a number of earning series available from ABS sources, and differences are observed when comparing these sources over time. Many factors contribute to the divergence in earnings, such as changes in wage rates, variations in hours worked, and changes in the composition of the employee work force.
The following sections provide a number of examples of the use of earnings statistics, namely: distributional and compositional analysis; gender comparisons; and wage movements.
Distributional and Compositional Analysis
Distributional and compositional analysis can help answer questions such as:
It is useful to examine the distribution of earnings to determine whether most employees receive earnings near the average, or whether a few highly paid employees increase average earnings. When analysing earnings data, which has a skewed distribution with a long right-tail, the median is a better indicator of central tendency than the mean. However, to derive a median value, earnings for each employee in the survey are needed, i.e. the whole distribution. Both the EEH and EEBTUM collections provide distributional data as standard outputs.
Mean earnings are usually higher than the median as the mean earnings are influenced by outliers (graph 1). Relatively small numbers of highly paid employees contribute more to the numerator when deriving the mean, which results in a higher average. Generally, the larger the gap between the mean and the median for a group of employees, the more uneven is the distribution of earnings for that group of employees, indicating that a greater proportion of employees have earnings at the lower end of the distribution.
The graph below shows the distribution of non-managerial adult hourly ordinary time earnings from EEH, May 2012 survey. EEH data are more robust for analysing the distribution of earnings, as information is collected from businesses (from their payroll) but at an individual employee level. However, the EEH survey (used in graph 1) only has a limited number of characteristics of employees.
Graph 1: TOTAL WEEKLY CASH EARNINGS, Adult full-time non-managerial employees: May 2012
Source: ABS data available on request, Survey of Employee Earnings and Hours, May 2012.
Weekly earnings are affected not only by changes in the rate of pay, but also by any changes in the composition of the Australian workforce, including:
EEBTUM data from August 2013 show there was a higher proportion of high earners in older age groups compared to younger age groups. The distribution of weekly earnings of employees in the age groups between 35 to 54 years were more skewed (i.e. wider gap between the mean and median), compared to those in the age groups between 15 to 24 or 25 to 34 years where the distribution is more equal (i.e. narrower gap between mean and median). The differences in the earnings distributions between younger and older groups can partly be explained by compositional differences between these two age groups.
A higher proportion of employees in the 35 to 44 and 45 to 54 year age groups work full-time in their main job. In August 2013, just over half (52%) of the employees in the 15 to 24 years age group worked full-time, whereas around three-quarters of employees in both the 35 to 44 and 45 to 54 year age groups worked full-time in their main job (73% and 72% respectively). A higher proportion of employees in the age group of 25 to 34 also work full-time (79%). This includes people who move to full-time work after completing their studies and, being a younger age group, tend to have less caring responsibilities (EEBTUM, August 2013).
The August 2013 data from EEBTUM also show that a far greater proportion of young employees were paid for few hours, 29% of employees (excluding OMIES) aged 15 to 24 years were paid for between 1 and 14 hours per week, compared with only 6% of employees (excluding OMIES) aged 25 to 54 years. This is a contributing factor towards the relatively lower weekly earnings in the 15 to 24 year age group. The middle age groups (those aged 35 to 44 and 45 to 54 years) have higher proportions of employees generally in higher skilled occupations, and are therefore higher paid. Over half of the employees in the Managers and Professionals major occupation groups are in the 35 to 54 years age group (54% and 57% respectively), resulting in higher median earnings for these age groups. Graph 2 below shows the mean and median earnings for the major occupation groups for August 2013.
Graph 2: EMPLOYEES IN A MAIN JOB(a), mean and median weekly earnings by occupation: August 2013
(a) Employees excluding OMIEs
Source: ABS data available on request, Survey of Employee Earnings, Benefits and Trade Union Membership, August 2013.
However caution should be exercised, as earnings estimates from EEBTUM are not as robust because they are reliant on respondents' (or another responsible adults’) accurate recall of their (pre-tax) earnings. Also, measures provided from EEBTUM do not separate ordinary time earnings from overtime earnings.
The earnings data collected by the ABS can to some extent support comparisons of earnings by gender. However careful consideration is needed, as many factors other than gender influence the observed differences in average earnings between males and females. These factors include labour market participation, hours worked, industry and occupation. Therefore the observed differences in earnings are generally a reflection of the differences in male and female working arrangements.
It may be necessary to analyse other data sources to get a more comprehensive picture of the composition of the workforce. The LFS provides more timely and robust information about the composition of the labour force, as the data are collected every month and from a larger sample of households. Therefore latest available data from the LFS has been used for analysis of compositional differences within the employed population in this section.
Generally, when looking at ABS statistics for average earnings, male employees earn higher weekly cash earnings than female employees. Much of the difference between earnings of different groups can be explained by a variety of factors including the variation of hours worked and the types of work done, e.g. different occupations or prevalence of part-time work. For example, LFS data shows that in April 2014, 83% of male employees worked full-time, while 54% of female employees were employed full-time. Females employed full-time usually worked fewer hours per week on average (40.8 hours) than males (44.6 hours), whereas females employed part-time usually worked 19.2 hours per week on average compared to males who usually worked 18.5 hours per week on average.
The distribution of weekly earnings are heavily influenced by the proportion of people employed part-time. For example, data from the February 2014 LFS shows that the major occupation groups Sales Workers, and Community and Personal Service Workers, had the majority of people employed part-time (56% and 51% respectively). These two major occupation groups also have a relatively high proportion of females. More than half (61%) of all Sales Workers were females, and 66% of those females worked part-time. Females also counted for the majority of Community and Personal Service Workers (68%), and of those females, 58% worked part-time. The earnings data from EEH, May 2012, shows that these two groups also had the lowest median weekly total cash earnings of all occupation groups, $504 and $636 respectively.
The occupation groups Professionals and Managers have higher proportions of people employed full-time and the highest median weekly earnings. Professionals had 89% of males and 66% of females employed full-time, and Managers had 93% of males and 76% of females employed full-time (LFS, February 2014). The median weekly total cash earnings for Professionals was $1353 and for Managers it was $1642 (EEH, May 2012).
LFS data from February 2014 shows that the vast majority of people employed as Machinery Operators and Drivers and Technicians and Trade Workers were male (92% and 86% respectively), and of these relatively few were employed part-time (14% of male Machinery Operators and Drivers and 9% of male Technicians and Trade Workers). These two occupation groups also had above average median weekly total cash earnings ($1098 and $1080 respectively) (EEH, May 2012).
Graph 3: MEDIAN WEEKLY TOTAL CASH EARNINGS FOR ALL EMPLOYEES, By occupation and by sex: May 2012
Source: ABS data available on request, Survey of Employee Earnings and Hours, May 2012.
The Accommodation and food services and Retail trade industries had the lowest levels of median weekly total cash earnings in May 2012 (EEH) ($455 and $590 respectively). LFS data from February 2014 shows that these two industries also have relatively high proportions of females (56% and 54% respectively) and a relatively high proportion of part-time employment. Retail trade had 48% of its workforce employed part-time, with 58% of females in this industry working part-time. Accommodation and food services had 57% of its employees working part-time with 62% of females in this industry working part-time.
The industry with the highest median earnings was Mining ($2250 - EEH, May 2012), where 84% of the workforce were males working full-time (LFS, February 2014).
Graph 4: MEDIAN WEEKLY TOTAL CASH EARNINGS FOR ALL EMPLOYEES, By industry and by sex: May 2012
Source: ABS data available on request, Survey of Employee Earnings and Hours, May 2012.
As described above, differences in earnings between males and females could be due to many factors, including different jobs within different occupations or industries, differences in full-time and part-time work, and also hours worked. Therefore as many factors as possible should be considered when analysing data.
Wage Movement Analysis
A key element in monitoring labour market and economic performance over time is examining changes in earnings. As earnings paid to employees represent a significant component of operating costs for businesses, changes in wages can highlight inflationary pressures facing businesses and/or impact on productivity. Changes in average earnings can also reflect the impact of the economic cycle on the labour market, or sectors within the labour market.
Up until recent times, WPI and AWE were both compiled on a quarterly basis, although AWE has recently changed to a biannual frequency with May 2012 being the last issue produced on a quarterly basis. Both WPI and AWE continue to be released in respect of May and November reference periods, and the common reference periods often lead to comparisons between the two series. Caution should be exercised when making such comparisons as differences in the purpose and design of the two collections means they will often respond differently to economic events.
Specifically, the WPI's focus on holding quality and quantity constant (to produce a measure of change in the price of a unit of labour) means it is affected solely by broad labour market influences on rates of pay. AWE will be affected by a more comprehensive set of economic factors. These include: changes in wages and salaries associated with individual performance; changes in employment that can affect the distribution of various types of employees between two periods (e.g. full-time vs part-time; higher paid vs lower paid) or changes in the pattern of hours worked (e.g. increase in total hours worked, increase in overtime hours). All these changes can influence changes in earnings between two periods to different degrees, and can result in different movements being observed for WPI and AWE. It is recommended that WPI be used to measure the change in the price of labour, or changes in wages over time, for the reasons described above.
Many factors contribute to the level and changes in earnings. These factors can be difficult to analyse independently, as most are inherent in the changes in employment patterns and composition, wage rates, hours worked and technological changes. Data gathered at the individual level, such as from the EEH and EEBTUM surveys, allow for compositional and distributional analysis, which makes it easier to try and account for the differences in employment patterns. The more factors which are taken into consideration when analysing data in general, the more robust such an analysis will be.
The various ABS sources of earnings information provide a wide range of data for a variety of purposes. Estimates from a given source may differ from estimates from other sources resulting from differences in scope, coverage and methodology. The decision on which data to draw on depends on the purpose and type of analysis to be undertaken.
The ABS encourages users to consider relevant factors in order to facilitate the most informed decision making.
More information on sources of earnings data, including conceptual or methodological differences, can be found in the Explanatory Notes of each publication, and in Labour Statistics: Concepts, Sources and Methods (cat. no. 6102.0.55.001).
For further information contact the Labour Market Statistics Section in Canberra on (02) 6252 7206 or email <firstname.lastname@example.org>.
This appendix provides a summary of the ABS data sources or publications about earnings and earnings-related data.